Testing consumer perceptions regarding the treatment variation in our study "Consumer Demand for Responsibly Produced Products: Evidence from a Natural Field Experiment in E-Commerce"

How to cite

Irlenbusch, Bernd; Conrads, Julian; Sarin, Maivand; Sliwka, Dirk. 2025. "Testing consumer perceptions regarding the treatment variation in our study "Consumer Demand for Responsibly Produced Products: Evidence from a Natural Field Experiment in E-Commerce"." AEA RCT Registry. AEARCTR-0017447. https://doi.org/10.1257/rct.17447-1.0

Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.

Bernd Irlenbusch* , Julian Conrads , Maivand Sarin , Dirk Sliwka

* Primary investigator

Registry ID
AEARCTR-0017447
DOI
https://doi.org/10.1257/rct.17447-1.0
First registered
Last updated
Status
Completed
Trial dates
2025-12-15 – 2025-12-19
Countries
Germany ()
JEL codes
C9, D12, M14
Registry record
http://www.socialscienceregistry.org/trials/17447

Abstract

In our previous field experiment (not preregistered) we study if consumers reward a
responsible production of products with higher demand. More precisely, we run a natural
field experiment in the online shop of a German kindergarten backpack producer by
simultaneously activating two versions of the shop and randomly directing visitors to one
of them (A/B-testing). We include and emphasize information about the backpacks’
socially and ecologically responsible production in the CSR treatment but leave any such
information out in the control treatment. We do not find a significant treatment effect on
sales.
Reviewers were interested in how far our treatment actually worked in the study above,
i.e., in how far the treatment moved customer beliefs about the sustainability of the
brand. Reviewers and the editor argued that assuming that people do indeed read the
information, this can fail either because (i) prior beliefs are already too high for the belief
shift to affect purchasing decisions, (ii) the information is already incorporated in beliefs
(i.e. not new) or (iii) the information is not credible to customers who are bombarded with
lots of similar for other brands which is often greenwashing. Or, because people do not
care about sustainability or CR. To adress these questions we designed an online survey
by showing participants (between-subject design) the webpages of our two treatments
and asked them about their beliefs about the brand.

Intervention

In the survey we will show each participants one version of the seller's two webpages -
one in which we include and emphasize information about the backpacks’ socially and
ecologically responsible production in the CSR treatment and one in which we leave any
such information out.

Experimental design

We will ask participants to carefully look at the webpage and subsequently answer some
questions about whether they know the backpacks of the brand, and to rate perception
questions as the following each on a five-items Likert-scale:
- the quality of the shown backpack is very good,
- the likelihood that the backpack would be very highly functional,
- the likelihood that the backpack is highly reliable,
- the brand appears an environmentally safe brand,
- the brand is an environmentally responsible brand,
- the brand is a sustainable brand,
- the brand’s backpack visually is appealing for kids,
- the backpack's design is attractive,
- I like the look and feel of the brand's backpacks,
- the brand’s backpacks are aesthetically pleasing for kids,
- the brand appears to produce under fair working conditions,
- I think that the workers are treated fairly who produce the brand’s backpacks,
- the brand provides a safe and secure working environment,
- I feel that the brand’s sustainability commitments are generally reliable,
- I feel that the brand’s sustainability performance is generally dependable,
- I feel that the brand’s sustainability argument is generally trustworthy.

We also ask them whether they would purchase the shown backpack if they were
searching for such a backpack for a child.

In order to be able to control for personal characteristics of the participants we will also
ask several question about how similar they would rate (on a 7 item Lickert scale)
themselves in relation to a person for whom it is important
- to avoid environmental harm,
- to protect the environment,
- to respect nature,
- that people have equal opportunities,
- to care about people who are worse off,
- to help others,
- to enjoy life,
- to be influential,
- to have money and possessions,
- to work hard and be ambitious.

Design details

Randomization method
Randomization is done by the computer when subjects log in.
Randomization unit
Individual
Clusters
200 per treatment, i.e., 400 in total We first registered the experiment with only 100 per treatment but a power analysis suggested to go for 200 per treatment.
Observations
200 per treatment, i.e., 400 in total
Arms
200 per treatment
Primary outcomes
Since our subjectpool comprises participants who are unlikely who have children in kindergarden age, we expect that few of them are familiar with the brand. This was also the case when we ran the original A/B testing experiment becaus at this time it was a new brand. We further expect that participants' perceptions of the backpacks offered on the CR webpage will be significantly higher with respect to the backpacks’ socially and ecologically responsible production compared to the neutral treatment. We do not expect differences between treatments in the perceptions of the attractiveness and the quality of the backpacks, and whether they are appealing for kids. We also do not expect differences between treatments in there willingness to purchase the shown backpack if they were searching for such a backpack for a child.

Cited by

No citation data available for this registration yet.

Keywords